How to Export Transcripts From Custom GPTs (OpenAI GPTs Tutorial)
HTML-код
- Опубликовано: 27 май 2024
- 📚 My Free Resource Hub & Skool Community: bit.ly/3uRIRB3 (Check “RUclips Resources” tab for any mentioned resources!)
🤝 Need AI Solutions Built? Work with me: bit.ly/3K3L4gN
📈 Find out how we help industry experts sign their first 5 AI Agency clients, guaranteed: bit.ly/skoolmain
In this video I show you How to Export Transcripts From Custom GPTs using the OpenAl Assistants API, then analyze each transcript using ChatGPT and Make.com. This includes learning how to export transcripts from any chatbot and to Airtable via ManyChat, Replit and Make.com.
Keeping track and analyzing your transcripts can be an extremely powerful tool fully capture the power of GPTs. By exporting your CustomGPTs chat history to Airtable, you can to access and extract key information from your custom GPTs, extract insights and drive growth for your business or your clients'.
Timestamps:
0:00 - Intro
1:03 - How it Works
2:50 - Replit Setup
5:10 - Airtable Setup
7:23 - Manychat Setup
8:38 - Make Setup
11:12 - Summary
14:14 - Testing System
Let me know if you have any issues below 😁👇🏽
📚 My Free Skool Community: bit.ly/3uRIRB3
🤝 Work With Me: www.morningside.ai/
📈 My AI Agency Accelerator: bit.ly/3wxLubP
Extensive research done here.Thank you Liam.
Unreal value once again mate!
Amazing video Liam, I can now pitch this to my clients and also working with some good developers to scale my AAA.
This video is raw gold liam. thank you so much. The value you are giving us for free is enormous.
What do you think of a video/tutorial explaining how we should handle the API and token usage (cost) of both Openai and Voiceflow after we sell the chatbot to a client?
Agreed. Advice about pricing our services against cost of tooling would be awesome.
yeah, AI automation tightly depends on the cost of deploying and maintaining the system. some simulation runs with cost outputs would be great. after all, why automate it when a human can do it manually cheaper?
Thank you Liam, great value.
Thank you, Liam - Please continue with this content , this is so valuable - this is the reason I am sub
It’s great to watch your videos, let’s learn together!Thank you❤
You work really hard Liam
Thank you, Liam! you. are the most valuable RUclipsr for me. This kind of content is the reason I am subscribed, keep going - you are highly recommended from me
This is an EXCELLENT video!!!
Thank u liam. u are a hero for us
Great video. Thanks for sharing this @Liam
Simply Legend❤
Really great one Liam! The tools section in the replit code is a bit complex. I mean could you please make a video on explaining how to change the components in the code for custom service chatbot, persona chatbot etc? You already made a lead gen chat bot. This is mainly to increase the knowledge on how to change the tools section of code while making different chatbots.
Btw I appreciate your work man. You always provide interesting content that is up to date. I mean the assistance API has just realised and you've made great content on how to make use of it.
Hey Liam huge thanks for this awesome video. Can this template only send an entire threat or can you also change it to send only certain ai responses within the threat to an external place. Can you also use this template to send an ai response to a google doc or a word doc instead of airtable?
Great video thanks a lot, maybe one small feedback, I appreciate your speaking speed, data flowing fast to my brain, but it does require max concentration, I wonder if the video editing effects that look great are always helping the transfer of data or are clogging up slightly our tiny human bandwidth, I think visual effects are great when they illustrate what is said so what we here and what we see syncs up seamlessly 😀
Hi Liam interesting video but have you got any idea how to connect messages from linkedin to gpt ?
Great
Hey Liam, great series! What are the GDPR implications on holding transcripts? Does the chatbot need to inform the user that the transcript and their details will be held?
No response on this question?
Great video btw liam, btw I have an idea pls tell me if its possible or not. Lets say I want to create an instagram lead qualifying bot which qualifies inbound leads and books them into appointment. I wanna do it this way, 1. Create a custom GPT (idk to code so, just the plain old simple custom gpt feature by OpenAI) I will prompt it like this you are an instagram lead qualifier you have to do this ask question and book them into an appointment. I will create custom GPT like that. 2. Then integrate the Custom GPT with voiceflow and then voiceflow will be integrated with manychat. The process will be like this once the user sends a trigger keyword, the process will be triggered. first trigger word from manychat to voiceflow then to Custom GPT then the GPT will greet and ask question and this is how I will be capturing the data. if the response from the GPT is a question like say what is your monthly revenue (example of a question) then I will set a capture data for these type of response from the GPT so it will capture the data which comes from the user through many chat as it voiceflow is an intermediate for manychat and the Custom GPT. would be really helpful if you could tell me if this will work. this seems like a lot of work but I think it is easy. Appreciate for posting invaluable content.
Why not just skip the Voiceflow step?
@@jacobporte3084 voiceflow has some features I can use it sure is a good idea.
hey how to do this airtable connection for website widget chatbot? can you make a video to explain it?
Hi Liam, great Video as always, i think your website is down at least the resource link is not loading for me.
Hey Liam! thank you for this AMAZING video, this is very important for us!
there is an Error with loading the "getSchema" function (according to Make) and this makes the task (connect make ) impossible, can you help us with that?
Hi, for a couple of days, I have also faced this issue. Make shows error- Failed to load data! Function 'getSchema' finished with error! Cannot read properties of undefined (reading 'fields').
Did you also see same error? Are you able to resolve your issue?
Thanks Liam trying to figure this out! When I press 'Run once' @make. I get this error. Do you know what this means? --> UDT (User defined type) not found: UDT not found 'json:19909'
I encountered the same issue too, unable to save the scenario as well
Good question, seems like the data structures don't carry across. You'll need to generate one yourself and save it. For the of the "JSON Parse" block you need to add a new data structure, to auto generate it you can run the HTTP node once, copy the output and use that to generate the output structure.
See the "sample data" box here: www.make.com/en/help/tools/data-structures#:~:text=Data%20structure%20generator,-Data%20structures%20don&text=You%20can%20make%20it%20easier,can%20then%20be%20manually%20modified.
@@LiamOttley I've got the same issue as well mate.
same issue plese give us totorial :) I dont understand what I must do
I've got the same issue please help us Liam :( so close to being ready to ship!
Hi, where is your replit template in description to download?
I can't find it
Thanks Liam! According to Airtable's website, its API keys will be deprecated by the end of January 2024. After this date, API keys will stop working... so Liam, what can I do with this?
Can you please update how this is done with Agentive?
do we have buy gpt 4 assistant api for 20$ or we get it for free?
someone, can explain me, how to deliver this type of service, how send the airtable part, can i ask business owner to sign in, the air table, and create fuction base on that, some explain me if you know.
hey liam , can you please help? according to Make there is an error with 'getSchema' func , can you chack this is out ?
Hi, for a couple of days, I have also faced this issue. Make shows error- Failed to load data! Function 'getSchema' finished with error! Cannot read properties of undefined (reading 'fields').
Did you also see same error? Are you able to resolve your issue?
I think this is because your database name or table name is not the same as Threads; I would make sure you're copying the exact template
Getting this error on Make when trying to link it with airtable Function 'getSchema' finished with error! Cannot read properties of undefined (reading 'fields'). :(
Which node is the error happening on?
Its happening on all three of my airtable nodes so Im assuming its something with my airtable, I generated a new Airtable key but same issue @@LiamOttley
Same problem .
I had this too but fixed it i guess. I don't know why, but there were 2 'bases' of Smith's Solar CRM and selected the second one. So at 'table' there were also 2 'threads' and selected the second one below. You then will see the correct output fields like in the video. Change this also in the other 2 Airtable records. But I have another error right now @@LiamOttley
still not solved please how? :)
Bro this is hella expensive to have isn't their a better way without all this because bruh
Sar mujhe maaf kar do aur mujhe nahin tha ke main galat tarike se ek ko istemal kar raha hun aur aapki message aaya to Maine Dekha Ki main aisi galti kar Gai Jo Kisi insan ko Nahin Karna chahie